Update of fecal markers of inflammation in inflammatory bowel disease
Dr Steven Leach, Westfield Research Laboratories, Level 2, Sydney Children's Hospital, High Street, Randwick, NSW 2031, Australia. Email: firstname.lastname@example.org
The diagnosis, prognosis, and assessment of disease activity of inflammatory bowel disease (IBD) require investigating clinical, radiological, and histological criteria, as well as serum inflammatory markers. However, a range of fecal inflammatory markers now appears to have the potential to greatly assist in these processes. Calprotectin, a prominent neutrophil protein, was identified two decades ago as a potentially revolutionary marker for IBD. Following this discovery, numerous additional markers, including S100A12, lactoferrin, and M2-pyruvate kinase, have also been suggested as novel markers of IBD. In the present study, we provide an up-to-date review of fecal markers of IBD, and further, provide a novel analysis of each of these fecal markers in severe ulcerative colitis and compare their expression pattern in contrast to calprotectin.
The inflammatory bowel diseases (IBD) are idiopathic, lifelong, chronic intestinal conditions characterized by periods of remission and recurrent relapses.1 The two main types of IBD, ulcerative colitis (UC) and Crohn's disease (CD), might be defined based upon endoscopic, histological, and radiological investigations, with histological findings of paramount importance.2 Typical to UC is superficial inflammation limited to the mucosa of the colon. In contrast, CD is characterized by discontinuous skip lesions occurring anywhere in the gastrointestinal tract, demonstrating transmural inflammation,3 and in approximately 35% of cases, non-caseating granulomas.4–6
Diagnosis, prognosis, assessment of disease activity, and severity, in addition to outcome of therapy, are aspects that continue to present challenges for physicians in the treatment of IBD. For each of these aspects, there is no universally-accepted test or examination.7 Intestinal inflammation is a primary criterion for differentiating IBD from other diseases, such as irritable bowel syndrome (IBS). However, acute intestinal inflammation might also be seen in conditions, such as infectious gastroenteritis.8 In the assessment of disease activity, and for the tailoring of therapy in IBD, the determination of inflammatory activity is critical. At present, accurate monitoring of intestinal inflammation relies upon both clinical indices (based upon symptoms and clinical examinations) and endoscopy, in conjunction with histological, radiological, or cross-sectional imaging techniques.9 Clinical indices tend to be too complex and time consuming for daily routine practice,10 and are hindered by inaccuracy due to subjective components. Endoscopy is costly, invasive, and has been associated with morbidity, and rarely, mortality.11
Clearly, a simple, reliable, reproducible, and non-invasive test, with the ability to differentiate IBD from other gastrointestinal conditions, such as IBS, and to monitor disease activity, would be of substantial clinical benefit.11 Fecal markers are a non-invasive way of objectively measuring intestinal inflammation, with the potential to play a primary or adjunctive role in the assessment of disease activity.12 Consequently, in recent decades, a number of fecal markers have been evaluated for their ability to differentiate and monitor IBD disease activity. While an ideal marker for IBD is yet to be identified, some interesting markers with significant potential have been evaluated.7 In this review, some of the most promising disease-specific markers, which have the potential to advance diagnostic and disease monitoring practices, will be discussed. Of particular interest are lactoferrin, M2-pyruvate kinase (PK), and two members of the S100 family of calcium-binding proteins: S100A12 and calprotectin.
Markers of intestinal inflammation
In patients with IBD, the presence of active gut inflammation is associated with an acute-phase reaction and the migration of leucocytes to the gut. As a corollary, a large number of acute-phase proteins are produced.7,9 In direct contact with the intestinal mucosa, the fecal stream should therefore contain specific markers of mucosal disease, consistent with the presence and degree of intestinal inflammation. Histological features of UC and CD include the aforementioned leucocyte infiltration, with subsequent sloughing of the cells and their products into the bowel lumen. Contemporaneously, the mucosa exhibits increased permeability and loss of normal barrier function.11 As a result of these processes, potential fecal biomarkers include the fecal excretion of leucocytes, leucocyte products, and serum proteins.12
Historically, the instability of inflammatory markers in the stool has led to difficulty in accurately assessing inflammatory products in the stool. The presence of fecal white cells can be a useful indicator of gut inflammation, after exclusion of infection. However, the accuracy of this marker relies upon prompt examination of the stool sample to avoid degradation of white and red blood cells by gut bacteria. Levels of α-1-antitrypsin can be increased in the stool consequent to the disruption of the intestinal barrier, but this marker has proven to be inaccurate, relating poorly to mucosal inflammation.13 Proteins released from neutrophil secretory granules, such as myeloperoxidase, have also been used as inflammatory markers for IBD. Myeloperoxidase is stable for at least 3 days in the stool,14 with several studies indicating that myeloperoxidase is elevated in IBD.14–16 However, there is significant overlap in myeloperoxidase levels in IBD compared to healthy controls. Masoodi et al. reports a sensitivity of 89% and specificity of 51% for fecal myeloperoxidase, distinguishing adult UC patients from aged-matched healthy controls.17 The limitation of myeloperoxidase as an inflammatory marker might be its cationic charge causing adhesion to fecal particles.18 When combined with inadequate fecal extraction techniques, accurate quantification of myeloperoxidase in the feces cannot be achieved. Therefore, to date, myeloperoxidase has shown to be of only limited utility as an inflammatory marker for IBD.
Currently, several standard, non-specific serum markers of inflammation are commonly used to aid in the diagnosis of IBD and the monitoring of IBD disease activity. These include the erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), acute phase protein (albumin), and platelets. Of these, ESR and CRP are the most commonly used; however, they have inadequate accuracy in identifying active gastrointestinal tract mucosal inflammation.1,19 Consequently, other more accurate markers of intestinal inflammation are required.
The human calcium-binding S100 proteins, a family of 21 proteins grouped within the larger EF-hand protein superfamily, are associated with a number of diseases.20 Three proteins, S100A8 (MRP8, calgranulin A), S100A9 (MRP14, calgranulin B), and S100A12 (calgranulin C), known as the calgranulin subfamily, have extracellular roles. These proteins act as markers of inflammation, and might also contribute to IBD pathogenesis.21
Calprotectin, a heterocomplex of S100A8 and S100A9, is a calcium-binding protein with antimicrobial protective properties derived predominately from neutrophils, and to a lesser extent, from monocytes and reactive macrophages.22 Calprotectin constitutes approximately 5% of the total protein and up to 60% of the cytosolic protein in human neutrophils.23 As such, the fecal calprotectin concentration is proportional to the influx of neutrophils into the intestinal tract, a hallmark of active IBD.7 Due to resistance to degradation, this marker has excellent stability in feces.24 Calprotectin levels correlate with 111Indium-labelled leucocyte excretion25 and intestinal permeability.7 There is no indication of gender differences in calprotectin levels in health or disease.2 It has been shown, however, that fecal calprotectin levels vary with age: newborn infants have high calprotectin levels, declining over the first weeks of life and reaching comparable levels to adults by 5 years of age.26–28 Furthermore, considerable day-to-day variation has been demonstrated in patients without colonic inflammation or neoplasm, suggesting that fecal calprotectin is influenced by factors other than disease.29
Fecal calprotectin levels have consistently been shown to be elevated in both adults and children with IBD relative to healthy controls.23 Fecal calprotectin also has a role in the investigation of symptomatic patients, in order to separate those with functional problems (such as IBS) from those with IBD. Tibble and colleagues30 were able to discriminate adults with active CD from those with IBS with 100% sensitivity and 97% specificity using a cut-off of 30 µg/g. Similar distinctions have also been made for pediatric patients with active IBD.31,32 von Roon and colleagues33 evaluated the diagnostic precision of fecal calprotectin for IBD using prospective studies comparing fecal calprotectin against the histological diagnosis, and calculated a sensitivity of 95% and a specificity of 91% for the diagnosis of IBD (vs non-IBD diagnoses). Furthermore, in a recent meta-analysis including 13 diagnostic accuracy studies, van Rheenan et al.34 demonstrated that fecal calprotectin is a useful screening tool for identifying those patients with suspected IBD who are most likely to need endoscopy. These investigators found that the use of fecal calprotectin as a screening test would result in a 67% reduction in the number of adults requiring endoscopy. Consequent to false-negative test results, this approach would also delay diagnosis in 6% of adults. Additionally, the specificity for calprotectin in the exclusion of IBD was found to be significantly better in studies of adults than in studies of children and teenagers. While fecal calprotectin is a good indicator of gut inflammation, levels are also elevated in other gastrointestinal disorders and during non-steroidal, anti-inflammatory drug use.22
Due to the correlation between fecal calprotectin and leucocyte excretion, fecal calprotectin levels are associated with the degree of IBD activity evaluated with clinical, endoscopic, and histological parameters.30–32,35,36 Interestingly, Bunn et al.31 report in a pediatric study that calprotectin concentrations correlate more closely with histological rather than endoscopic findings. These findings suggest that fecal calprotectin might be more sensible than endoscopy in evaluating IBD activity. Furthermore, Canani et al.1 demonstrated that fecal calprotectin levels show a strong relationship with the degree of mucosal inflammation in a group of children with known IBD.
As demonstrated by Costa and colleagues,19 calprotectin determination appears to better reflect disease activity in UC than in CD. A lack of correlation has been shown between fecal calprotectin levels and the Pediatric CD Activity Index.37 Both this score and the CD Activity Index might not be sufficiently-sensitive tools to reflect subclinical inflammatory activity present in CD.38 As such, it has been suggested that stratification based on phenotypical pattern (inflammatory, structuring, or fistulizing) could improve the predictive capacity of calprotectin in CD.22
Typically, patients with IBD experience periods of remission, with intermittent relapses characterized by increased intestinal inflammation. As the timing of these relapses is unpredictable, monitoring of disease remission has traditionally been performed with the aid of clinical symptoms. However, because these symptoms do not typically manifest early when inflammation is minimal, most flare-ups only come to medical attention after the inflammatory response has become well established.23 One of the most promising aspects of fecal calprotectin is its potential to predict relapse in IBD.19,39 Fecal calprotectin normalizes with microscopic mucosal healing.40,41 The elevation of the fecal calprotectin level in patients with IBD in remission, however, is associated with a higher risk of clinical relapse.19,31 Tibble et al.,39 in a pioneering study, demonstrated that elevated fecal calprotectin levels were associated with a 13-fold increased risk for relapse. In their study, 43 patients with CD and 37 patients with UC, in clinical remission, were assessed for 1–4 months using clinical disease activity indices with fecal calprotectin measured. Following a 12-month follow up, the sensitivity and specificity of calprotectin (with a cut-off of 50 µg/g) for predicting relapse in all patients with IBD were 90% and 83%, respectively. Although these values were similar when patients with CD or UC were examined separately, a later study by Costa and colleagues19 was unable to demonstrate a significant increased risk of relapse in patients with CD in remission. They concluded that fecal calprotectin predicts relapse more accurately in UC only.
S100A12, also known as EN-RAGE (extracellular, newly-identified receptor for advanced glycation end-products) and calgranulin C, is expressed as a cytoplasmic protein in neutrophils, and has pro-inflammatory properties, including potent chemotactic activity, comparable with other chemotactic agents.42,43 A ligand of RAGE, S100A12 likely activates the nuclear factor-κB signal transduction pathway.44 Tumor necrosis factor-α, a cytokine upregulated by nuclear factor-κB activation, further enhances S100A12 expression.44 These properties are most relevant to IBD,2 with infiltration of S100A12-positive polymorphonuclear cells potentially contributing to the invasion of other leucocytes.42 This may suggest that S100A12 contributes to the processes of gut inflammation and furthermore that S100A12 levels might reflect the presence and severity of intestinal inflammation.45 Serum S100A12 levels correlate with fecal levels.45 Furthermore, S100A12 has been shown to be evenly distributed throughout feces and is temperature stable for 7 days; characteristics desirable for a non-invasive, stool-based disease marker. As with calprotectin, there appears to be no gender difference in fecal S100A12 levels.45
It has previously been demonstrated that serum S100A12 is elevated in inflammatory disorders, such as rheumatoid arthritis46,47 and cystic fibrosis.48 More recently, fecal S100A12 has been shown to distinguish, with high sensitivity and specificity, chronic IBD from healthy individuals and/or from non-organic disease, including IBS.49 In a study by de Jong et al.,45 stools were collected from 25 healthy patients with no gastrointestinal symptoms and 23 children with newly-diagnosed IBD at the time of diagnosis and during treatment for IBD. Fecal S100A12 distinguished children with active IBD from healthy controls with a sensitivity of 96% and a specificity of 92% using an immunoassay with a cut-off of 10 mg/kg. Fecal S100A12 was elevated at diagnosis and fell during therapy in children entering clinical and biological remission with normal CRP levels. S100A12 levels correlated with disease activity scores for children with pancolitis. While these results suggested that measuring S100A12 is not a viable alternative to colonoscopy in IBD diagnosis, it might be valuable as a screening tool and for monitoring disease activity.
A further study expanded upon these findings.37 Sixty-one children presenting with intestinal symptoms were enrolled prospectively, with fecal S100A12 and calprotectin measured. Thirty one of these children were shown to have IBD on subsequent tests. In these children, both fecal S100A12 (median: 55.2 mg/kg) and calprotectin (median: 1265 mg/kg) were elevated compared to those children without IBD (n = 30, S100A12 median: 1.1 mg/kg, P < 0.0001; calprotectin median: 30.5 mg/kg, P < 0.0001). Upon further analysis, using a cut-off of 10 mg/kg, S100A12 gave a sensitivity and specificity of 97%, respectively, for the detection of IBD. However, a 50 mg/kg cut-off for calprotectin yielded a sensitivity of 100% and specificity of 67%, lower than the specificities reported by other studies.50 Both fecal markers were superior compared to the sensitivities and specificities of any standard inflammatory test in this population.37
Several studies have also demonstrated a role for S100A12 in the adult setting. Foell et al.35 demonstrated correlations between serum S100A12 and disease activity, and showed that serum levels fell after intervention with infliximab. Subsequent studies have also shown elevated serum S100A12 in IBD, however, with only modest sensitivities and specificities in distinguishing IBD and non-IBD patients.21,51 Following earlier work examining serum levels of S100A12 in the context of IBD, Kaiser et al.49 illustrated that fecal S100A12 levels could distinguish IBD from IBS, with 86% sensitivity and 96% specificity, and to also differentiate IBD from normal controls, with 86% sensitivity and 100% specificity. Their study also demonstrated that S100A12 was superior to calprotectin or other biomarkers in correlation, with an inflammatory score incorporating endoscopic and histological findings.49 The role of S100A12 as a marker of future relapse has not yet been considered in pediatric or adult settings. Prospective studies are required to elucidate this potential role.
Lactoferrin is an iron-binding glycoprotein identified in the secretions overlying most mucosal surfaces that interact directly with external pathogens, including saliva, tears, vaginal secretions, feces, synovial fluid, and mammalian breast milk.52–54 Lactoferrin is a major component of the secondary granules of polymorphonuclear neutrophils and is shown to be a primary factor in the acute inflammatory response.54,55 In the intestinal lumen, fecal lactoferrin levels quickly increase with the influx of neutrophils during inflammation.12 Lactoferrin has antibacterial activity and is resistant to proteolysis in the feces;56 it is unaffected by multiple freeze/thaw cycles54 and might remain stable in stool for as long as 5 days, compared to 7 days for calprotectin.24,57 Following storage at room temperature for 48 h, fecal concentrations of lactoferrin were 90% of initial levels, contrasting with fecal concentrations of calprotectin being 82%.23
Previous investigations of infectious diarrhea, using a now widely-available commercial ELISA, showed that fecal lactoferrin is highly sensitive for detecting fecal neutrophil infiltration.57 More recently, its utility as a sensitive and specific marker of intestinal inflammation in patients with chronic intestinal disease has been investigated.9,23,24,36,54,58–61 Several studies have indicated the usefulness of measuring lactoferrin in patients with IBD. For example, Sugi and colleagues57 investigated lactoferrin, polymorphonuclear neutrophil (PMN) elastase, and lysozyme together with myeloperoxidase in fecal material and whole-gut lavage fluid from IBD patients. They concluded that lactoferrin was superior as a marker of intestinal inflammation. In contrast, Silberer et al.62 reported that only PMN elastase and calprotectin, but not lactoferrin, myeloperoxidase, or lysozyme, were able to differentiate chronic IBD from IBS, and were correlated with severity of inflammation, as determined by ileocolonoscopy. In an another study, lactoferrin and calprotectin were shown to differentiate active IBD from inactive IBD and IBS in 80% of cases, compared to 74% for PMN elastase and 64% for CRP.9 Following these results, the authors suggested that using all three markers (lactoferrin, calprotectin, and PMN elastase) in a composite index might be an additional non-invasive tool for the management of patients with UC. In a further study of 215 patients, including 184 with known IBD, Kane et al.54 demonstrated that fecal lactoferrin was 86% sensitive and 100% specific in distinguishing IBD compared to healthy controls and patients with IBS. Significant differences were also seen between patients with active and inactive IBD, and those with inactive IBD had increased fecal lactoferrin levels compared to healthy controls and patients with IBS. Furthermore, fecal lactoferrin concentrations were noted to differ between patients with CD and those with UC. In contrast, another study found that fecal lactoferrin was unable to distinguish between active CD and UC.58 Using a rapid fecal latex agglutination test for the detection of lactoferrin, Fine and colleagues63 demonstrated that fecal lactoferrin was 90% sensitive in detecting inflammation in IBD and had a negative predictive value of 99%. Additionally, a recent pediatric study suggested that it is a sensitive and specific marker of inflammation in children with IBD, with the level correlating well with both clinical disease activity indices and ESR.24 Finally, the potential utility of fecal lactoferrin measurements in predicting patients who are at risk of relapse has been investigated.61 Although only performed with a small sample size, the results obtained in this investigation were promising, raising the possibility that elevations in lactoferrin might presage clinical flares.
PK is a key enzyme in the glycolytic pathway and is expressed by all cells.64 In humans, PK exists in dimeric and tetrameric isotypes.65 The tetrameric form (M1) has been found in skeletal muscles, the heart, and brain, while the dimeric M2 type (termed M2-PK) is present in undifferentiated and proliferating tissues and can be reliably detected in serum and feces.65,66 Increased PK activity in polymorphonuclear neutrophils is seen in patients with polytrauma,67 chronic cardiac failure,68 gastrointestinal tumors,69 and more recently, in pouchitis.66 However, the role of M2-PK in intestinal inflammation is not known. Active IBD is intrinsically linked with increased cell turnover and rapid division; cell turnover returns to normal once inflammation has resolved.70 As such, it has been postulated that fecal concentrations of M2-PK would be elevated in patients with IBD.64 Given this hypothesis and that M2-PK is a useful marker for gastrointestinal cancers (73% sensitivity and 78% specificity)69 and pouchitis,66 fecal M2-PK has also been investigated as a novel, potentially valuable, non-invasive marker of disease activity in IBD.64,65 The enzyme is stable for 2 days at room temperature, and the ELISA test can be readily carried out in a routine laboratory.69
In one particular study, Chung-Faye et al.64 obtained fecal M2-PK and calprotectin measurements from 148 patients: 50 with UC, 31 with CD, 43 with IBS, seven with colorectal carcinoma, and 17 with miscellaneous conditions (excluded from analysis). It was found that median M2-PK values were significantly elevated in UC, CD, and colorectal carcinoma compared to IBS, with a strong linear correlation of M2-PK, with calprotectin levels demonstrated. Using a predetermined cut-off level for normal fecal M2-PK, a sensitivity, specificity, and positive predictive value of 73%, 74%, and 89%, respectively, for differentiating organic disease from IBS was obtained. Furthermore, M2-PK levels (U/mL) were shown to be significantly elevated in active, compared to inactive, IBD (CD: 30 vs 0.55 U/mL, P < 0.005; UC: 40 vs 1.2 U/mL, P = 0.006). Fecal M2-PK levels have also been shown to be significantly elevated in children with IBD, with levels in UC in remission being significantly lower than in active disease.65
Being only a relatively new candidate marker of IBD disease activity, further studies are needed in order to properly assess the clinical value of fecal M2-PK in diagnostic work-up, screening, and treatment.65 Nevertheless, preliminary investigations suggest that M2-PK is a sensitive and relatively specific marker for organic gastrointestinal pathology in both adults and children.64,65 Whether M2-PK can be used to predict relapse in asymptomatic IBD patients is yet to be tested.64
Comparison of fecal M2-PK with other biomarkers in acute, severe UC
Recently, Turner and colleagues71 presented the first systematic head-to-head comparison of calprotectin, S100A12, lactoferrin, and M2-PK in severe, acute UC. Incorporated within a large, prospective, multicentre cohort study assessing the outcome of severe pediatric UC,72 this substudy included baseline samples from 101 children (13.3 ± 3.6 years; Pediatric UC Activity Index [PUCAI] at admission 72 ± 12 points), with disease severe enough to trigger intravenous corticosteroid treatment. Of the 101 children, 26 (26%) eventually failed steroid treatment and required salvage therapy by discharge. Analysis was conducted to elucidate the ability of the four markers to measure response to treatment, and to predict steroid refractoriness and outcome. Median values at baseline were elevated for all four markers. However, none of the markers were able to measure response to treatment in severe UC. Interestingly, however, M2-PK was found to have a good predictive validity to identify those failing intravenous steroid treatment, although less than the PUCAI, suggesting its usefulness as an objective measure for disease activity and for predicting treatment outcome in the severe UC setting. In comparison, fecal calprotectin had a fair predictive validity, whereas S100A12 and lactoferrin had none.
With the authors' permission, Spearman correlation analyses were performed for every marker combination using data procured from the study. Calprotectin and lactoferrin were found to correlate well, whereas the remaining combinations demonstrated considerably weaker correlation (Table 1). The good correspondence between calprotectin and lactoferrin might suggest a degree of concordance in their expression patterns. While this would suggest little value in pairing calprotectin with lactoferrin, simultaneously measuring calprotectin or lactoferrin together with S100A12 and M2-PK could prove beneficial.
Table 1. Correlation of fecal inflammatory markers
Endoscopic assessment, the current gold standard for the diagnosis, assessment, and monitoring of disease activity in patients with IBD, is overly complex, time consuming, costly, invasive, and at times, dangerous. Fecal biomarkers promise to significantly alter the way in which IBD is diagnosed and managed.11 While it is unlikely that they will ever replace invasive tests, such as endoscopy, which will always be necessary for definitive tissue diagnosis, fecal markers could be useful in reducing unnecessary invasive investigations.24,34 However, clearly, much work remains to be done. The currently-available fecal biomarkers allow the non-invasive assessment of specific aspects of gut inflammation. Although various roles have been established, none of the current markers are useful in all clinical settings. Further work is required to more fully define the roles of these markers. Nonetheless, there is clearly the opportunity to incorporate one or more of these markers into standard clinical practice for the routine assessment and monitoring of IBD.